Ecosystem stability is characterized by multidimensionality, hierarchical organization, and scale dependence. The lack of a clear conceptual framework has often led to confusion. We systematically examined the connotations, core dimensions, organizational levels, and spatiotemporal scale dependencies of ecosystem stability. In a narrow sense, ecosystem stability commonly refers to temporal stability of productivity, which is typically quantified as the ratio of the mean to the standard deviation of productivity (the inverse of the coefficient of variation). In a broader sense, ecosystem stability represents the overall capacity of ecosystems to maintain key structure, function, and ecosystem service provision under fluctuated environmental conditions, external disturbances, and long-term changes. It encompasses, at least, several dimensions, including resistance, recovery, resilience, persistence, invariability, and multi-stability. Based on the broad conceptual framework of ecosystem stability, we reviewed the definitions and major research advances from three perspectives: structure and function, different levels of ecological organization, and spatiotemporal characteristics. We further summarized and clarified the multidimensional and multilevel conceptual system of ecosystem stability. Current studies on ecosystem stability still face several major challenges: diverse but inconsistent definitions, numerous indicators lacking integration, a stronger emphasis on functional stability than structural stability, and insufficient understanding of cross-scale stability mechanisms due to the prevalence of single-scale studies. Future research should establish a unified conceptual framework and standardized indicator system for ecosystem stability, while further strengthening investigations into structure-function coupling, cross-level transmission mechanisms among ecological organizational hierarchies, relationships across spatiotemporal scales, and comprehensive assessments of multidimensional ecosystem stability.

Wetlands support rich biodiversity and provide unique habitat functions, playing a critical role in maintaining regional biodiversity and ecological processes. Evaluating the stability of habitat function is fundamental for identifying the quality and health status of wetlands and for guiding wetland conservation, restoration, and sustainable management. However, available habitat function assessments have mainly focused on individual indicators, lacking systematic multidimensional indicator frameworks and integrated evaluation approaches. To address this knowledge gap, we proposed a systematic methodological framework for assessing the stability of wetland habitat function. First, an evaluation indicator system comprising 11 sub-criteria layers and 43 candidate indicators was developed from four criterion layers, including environmental factors, biological composition, landscape patterns, and stress disturbances and conservation management. According to the temporal characteristics, those indicators were further classified into time-series ones and structural-state ones. Based on the constructed annual habitat function index (HFI) time series, stability was quantified from three dimensions: variability, resistance, and recovery. Specifically, the coefficient of variation was used to characterize long-term fluctuations. The relative change in HFI before and after disturbance was used to quantify resistance. Recovery rate, recovery magnitude, and recovery time were employed to describe recovery capacity. Furthermore, two approaches, including dimension-specific representation and a comprehensive habitat function stability index (HFSI), were established for integrated stability assessment. This framework systematically incorporated ecosystem stability theory into wetland habitat function evaluation, providing a quantitative basis for identifying wetland degradation, implementing conservation and restoration interventions, and supporting adaptive management.

Coastal wetlands are unique ecosystems formed by the interaction between land and sea, serving as important ecological barriers supporting sustainable development. They possess high biological productivity, strong ecological resilience, and significant carbon sink functions, but are currently under dual pressures from intense human activities and global climate change. Here, I systematically sorted out the development context and trends of research on coastal wetlands’ response to global change. The global research on coastal wetlands has achieved multiple key transformations. At the cognitive level, it has shifted from a die-off prediction of coastal wetlands under climate stress to a resilience cognition that emphasizes active ecological adaptation. In mechanism research, it has expanded from focusing on regional processes to basin-scale coupled intervention. The concept of conservation and restoration has upgraded from “area priority” to “quality orientation”. The realization of ecological value has moved from the scientific cognition level to market-oriented practice. These transformations indicate that the research, conservation and restoration of coastal wetlands have entered a new stage of development. Meanwhile, this field still faces core challenges, including insufficient analysis of multi-interface ecological processes and driving mechanisms, imperfect precise simulation and decision support systems, and immature technologies for ecological function improvement. In the future, the research, conservation, and restoration of coastal wetlands should adhere to the principles of land-sea coordination and sustainable development, with the core goals of enhancing ecological resilience, maximizing ecological functions, and promoting the marketization of ecological value. It is necessary to focus on deepening the analysis of multi-interface processes and mechanisms, constructing a precise simulation and decision support system, developing near-natural ecological restoration and synergistic efficiency technologies, improving the ecological product trading system, and building a land-sea coordinated protection pattern to realize the synergistic governance of river basins and coastal zones.

Wetland vegetation regulates surface energy balance through biophysical processes and modulates land surface temperature. However, a systematic understanding of such effect remains insufficient. We systematically reviewed the effects of wetland vegetation changes on land surface temperature from three perspectives: distribution patterns, vegetation coverage, and phenological dynamics. The regulation of key biophysical processes such as evapotranspiration, surface albedo, and turbulent exchange by wetland vegetation changes exhibits pronounced spatiotemporal heterogeneity, thereby leading to divergent surface energy partitioning and temperature feedbacks. Changes in vegetation distribution patterns can alter surface roughness and evaporation potential, thereby modifying the spatial configuration of turbulent exchange and water-heat fluxes, with consequences on land surface temperature. Increases in vegetation coverage drive daytime cooling by enhancing evapotranspiration, but this effect displays marked spatial variability. For example, it is more pronounced in semi-arid regions than in humid ones, and reverses to a warming effect during winter at high latitudes. Seasonal variations in the physiological activity of wetland vegetation can regulate the intensity of water vapor exchange, alter the temporal allocation of surface energy, and thus drive seasonal fluctuations in land surface temperature. There are some limitations of current research: inadequate differentiation and mechanistic quantification of the coupled interactions between wetland vegetation dynamics and hydrological processes, and insufficient process-based attribution of energy balance components, particularly latent, sensible and ground heat fluxes. The improved parameterization and ground validation of models that simulate wetland vegetation-hydrology-energy exchange processes are critically needed. In the future, it is necessary to integrate multi-source data and models, quantitatively separate the contributions of radiative and non-radiative processes, and couple biogeochemical processes to construct a unified framework for assessing climate effects, which will provide theoretical support for wetland conservation and restoration and accurate prediction of wetland climate-regulation functions under the background of global change.

As typical land-sea transition ecosystems, coastal wetlands are of great significance in maintaining ecological security and promoting regional sustainable development. The conservation and management of coastal wetlands is currently in the transition stage from rescue restoration to systematic governance in China. Understanding the practical lessons and experiences of international wetland protection and restoration practices is a key step to facilitate the sustainable use of wetlands in China. We systematically investigated the international management cases of coastal wetland that are similar to China’s coastal wetlands in terms of location, environment, development stage, as well as challenges. We further reviewed the governance objectives from the aspects of system construction, capital guarantee, management means, community participation, and international cooperation, and refined their common characteristics. Combined with the practical problems and basis in the practice of coastal wetland governance in China, we analyzed the applicable scenarios and localization paths of international experience, and put forward countermeasures and suggestions to improve the comprehensive management efficiency of coastal wetlands from the dimensions of legal system construction, space management, organizational coordination, scientific and technological support, capital guarantee, and multiple co-governance.

Wetlands are highly sensitive to hydrological disturbances. The carbon source-sink dynamics of wetlands are jointly regulated by hydrological regimes and soil structural heterogeneity. Under frequent water-level fluctuations and wetting-drying cycles, complex pore architectures and heterogeneous flow pathways are gradually deve-loped in wetland soils, generating preferential flow processes that substantially deviate from the assumption of homogeneous continuum media. As rapid conduits for water and solute transport, preferential flows can modify soil carbon transformation by regulating the redistribution of water, the transport of oxygen, and the availability of carbon substrates within heterogeneous soils. We systematically reviewed the physical mechanisms, observation techniques, and simulation methods of soil preferential flow formation in wetlands, and analyzed the impact of preferential flow on carbon transformation by regulating water redistribution and oxygen transport. By integrating typical scenarios of tidal dynamics in coastal zones and water level fluctuations in riparian zones, we elucidated biogeochemical processes driven by preferential flow and highlighted the amplification effects of non-steady flow processes under extreme events. Future research should focus on hydro-biogeochemical coupling mechanisms in wetlands, advancing the synergistic development of multi-scale observations and model simulations to deepen understanding of carbon transformation processes in wetland soils under preferential flow influence. This would provide theoretical support for precise assessment of wetland carbon sequestration functions and the formulation of emission reduction strategies under climate change.

Wetlands are critical ecosystems that support hydrological and climate regulation, biodiversity conservation, and ecological security. However, under the combined pressures of climate change and human activities, global wetlands are increasingly facing severe area loss, functional degradation, and habitat fragmentation. With the advantages including long-term time-series records, broad spatial coverage, and multi-source observations, remote sensing big data provide essential technical support for identifying wetland degradation, monitoring dynamic changes, tracking restoration processes, and enabling risk early warning. We systematically reviewed the current applications, technical approaches, assessment frameworks, and future development directions of remote sensing big data in wetland monitoring. First, we sorted out the characteristics of remote sensing big data sources and their applicability in wetland monitoring. Second, we comparatively analyzed recent advances in wetland information extraction techniques based on spectral features, temporal characteristics, machine learning, deep learning, and multi-source data fusion. Third, from the perspectives of hydrological processes, vegetation function, landscape structure, and temporal stability, we constructed a remote sensing-based assessment framework for monitoring wetland degradation and restoration, and explored its applications in degradation diagnosis, restoration trajectory tracking, effectiveness evaluation, and instability early warning. Finally, we identified key challenges in current applications, including difficulties in integrating massive heterogeneous datasets, limitations in spatiotemporal resolution, the complexity of wetland surface conditions, and insufficient ecological interpretation. Accordingly, we proposed future development pathways involving multi-source sensor synergy, model construction for complex scenarios, multi-factor environmental assessment, and innovation in governance models. This review would provide theoretical references and technical guidance for whole-process monitoring of wetland degradation and restoration, evaluation of ecological conservation and restoration, and intelligent wetland governance.

The sediment-water interface is a key zone for the release, retention, and transformation of internal phosphorus (P) in wetlands. Wetland vegetation can alter the source-sink status of P at this interface by modifying hydrodynamic conditions, rhizosphere redox environments, plant uptake, and microbial processes. We synthesized the mechanisms by which wetland vegetation regulates P transport and transformation across the sediment-water interface, including physical sediment stabilization, radial oxygen loss, and iron-phosphorus coupling, plant-microbe interactions, and litter decomposition feedbacks. We further compared the regulatory effects of different plant life forms and community characteristics. Emergent plants generally exhibit strong rhizosphere oxygenation and sediment-stabilizing capacities, thereby favoring P retention at the interface. Submerged plants can reduce P release by increasing near-bottom dissolved oxygen, taking up P, and suppressing sediment resuspension. By contrast, the long-term effects of floating-leaved and free-floating plants depend more strongly on vegetation coverage, litter management, and bottom-water oxygen conditions. The effectiveness of vegetation-mediated P regulation is jointly constrained by hydrological fluctuations and wetting-drying cycles, temperature, dissolved oxygen, pH, sediment properties, initial P speciation, and external P loading. During the growing season, plant P uptake and interfacial P retention generally predominate, whereas senescence or disturbance may increase P release risks through litter return, oxygen depletion, and weakened sediment stabilization. Consequently, the long-term effect of wetland vegetation on P cycling is strongly context dependent. Future research should focus on identifying thresholds under multiple interacting factors, elucidating functional plant-microbe interactions, developing high-resolution in situ monitoring approaches, and evaluating the long-term stability of P source-sink dynamics, thereby providing a scientific basis for wetland vegetation restoration and precise management of internal P loading.

Coastal wetlands are crucial “blue carbon” ecosystems. Previous studies primarily focus on vertical carbon sequestration in carbon sink assessments, often overlooking the key pathway of lateral carbon exchange. We systematically reviewed the biogeochemical transformation processes, key measurement methods, and driving mechanisms of key components in lateral carbon exchange (dissolved inorganic carbon, dissolved organic carbon, and particulate organic carbon) in coastal wetlands. Compared to vertical carbon burial rates, lateral carbon export rates in coastal wetlands are higher, with no differences among different types of coastal wetlands. The key processes include porewater exchange driven by tidal pumping and submarine groundwater discharge (SGD). Tidal processes drive periodic porewater exchange between sediments and overlying water bodies, promoting the release of dissolved substances from sediments. Meanwhile, multiple geophysical processes, including hydraulic gradients, tidal-induced seabed pressure variations, wave action, and density differences between seawater and groundwater collectively govern the occurrence and intensity of SGD. In terms of driving mechanisms, the tidal pumping effect, porewater exchange, and SGD are the core drivers of lateral carbon export, while extreme weather events (e.g., typhoons, heavy rainfall) dominate short-term carbon flux variations through triggering effects and hydrodynamic intensification. Carbon exported laterally to offshore areas achieves long-term sequestration via pathways such as the carbonate system buffering and sedimentation into deep-sea sediments, yet remains strongly perturbed by human activities and climate change. Future research should conduct fine-scale quantification of the physical-biogeochemical coupling processes in lateral carbon export, perform coordinated observations of multiple carbon components and isotopic tracing, and develop dynamic predictive models of carbon budgets under multiple environmental stressors, providing scientific support for precise global blue carbon accounting and adaptive coastal zone management.

Marshes are characterized by strong organic carbon accumulation capacity and important carbon sink functions. However, large-scale marshland reclamation for cropland can lead to substantial depletion of marsh soil carbon pools and impairment of soil organic carbon (SOC) stability. Iron minerals are important carriers of SOC and can form carbon-iron complexes (Fe-OC) through adsorption and co-precipitation, thereby contributing to SOC retention. Here, we summarize the effects of reclamation on SOC content and fractions in marsh soils. We further review how iron mineral forms, Fe-OC association mechanisms, organic carbon molecular composition, microbial processes, and environmental factors regulate the carbon-sequestration function of iron minerals, and evaluate the effects of reclamation on this function in marsh soils. We summarize current knowledge gaps and highlight the need to clarify how reclamation-induced changes in carbon sources affect the stability and thresholds of iron mineral-associated carbon sequestration, to elucidate the succession of Fe-reducing microbial communities and their roles in regulating iron mineral carbon sequestration, and to develop sustainable carbon management strategies for reclaimed marsh soils. This review would provid a better understanding of the mechanisms by which reclamation affects the stability of marsh soil carbon pools and provides a scientific basis for sustainable marsh management.

Wetland soil organic carbon comprises organic fractions that differ in biological origin, decomposition stage, and stabilization status. The preservation of plant-derived residue carbon and the accumulation of microbial necromass carbon represent two important pathways for SOC formation and stabilization in wetlands. Lignin phenols and amino sugars are commonly used to characterize lignin-derived plant residue components and estimate microbial necromass carbon, respectively. We synthesized the contrasting patterns of these two source-related residue carbon components across different wetland types and clarified the major drivers. Their contents vary substantially among wetland types, but the differences cannot be attributed to any single environmental factor. Instead, they emerge from the combined effects of plant inputs, microbial processing, decomposition and mineralization losses, and physicochemical protection. Under low-temperature, persistently anoxic, and strongly peat-forming conditions, the decomposition of plant residues is constrained, thereby promoting their preservation. Meanwhile, continuous plant inputs provide substrates for microbial assimilation and subsequent necromass formation. Water-table fluctuations, alternating redox conditions, sedimentation, and anthropogenic management can modify these processes and thereby regulate the relative contributions of plant-derived residue carbon and microbial necromass carbon. Future research should integrate multiple biomarkers with complementary multi-proxy approaches to elucidate the formation, transformation, and stabilization pathways of biologically derived residue carbon in wetlands, thereby advancing our understanding of wetland carbon pool formation and its responses and feedbacks to climate change.

Lake terrestrialization is a distinctive successional phase through which open-water ecosystems gradually transition into marsh-dominated wetlands. In floating marshes, intertwined roots and rhizomes form buoyant, peat-like mats that underpin marsh formation and persistence while serving as important interfaces for carbon input, transformation, and stabilization. Compared with typical peatlands and freshwater marshes, the input and stabilization mechanisms of root-derived carbon in floating marshes remain poorly understood. We reviewed current know-ledge of floating-marsh root traits and ecological adaptations and analyzed major root-derived carbon inputs arising from fine-root, coarse-root, and rhizome turnover, belowground residue accumulation, and rhizodeposition. We further highlighted three major stabilization mechanisms: 1) the enzyme-latch mechanism, driven by phenol oxidase inhibition and phenolic accumulation under anoxic peat conditions, together with the iron-gate mechanism associa-ted with iron redox cycling; 2) radial oxygen loss-mediated rhizosphere iron plaque formation and the associated “rusty sink” process; and 3) microbial transformation of root exudates into metabolic products and necromass, with implications for rhizosphere priming and microbial entombment. This review would provide a theoretical basis for evaluating the carbon sink function of floating marshes.

Soil is one of the largest carbon reservoirs in terrestrial ecosystems. Soil carbon sequestration plays a critical role in mitigating global climate change. Exogenous carbon inputs are key drivers regulating soil organic carbon mineralization and stabilization processes, which are widely present in both natural ecosystems and agricultural management practices. However, the effects of exogenous carbon inputs on soil carbon budget remain highly uncertain. Such exogenous carbon inputs may enhance carbon accumulation, but can also suppress or accelerate the decomposition of soil organic carbon through priming effects. We systematically synthesized the major mechanisms by which exogenous carbon inputs influence soil carbon turnover and stability, focused on the transformation and allocation pathways of exogenous carbon in soil, as well as the formation efficiency. Furthermore, we summarized the key regulatory factors, including microbial processes, mineral protection, and soil physical mechanisms. On this basis, we discussed the main controversies and limitations in current research and proposed future research directions, aiming to provide a theoretical foundation for enhancing soil carbon sink capacity.

Traditional carbon cycle research in estuarine wetlands has mostly been confined to static observations of interannual total precipitation and total carbon, struggling to reveal the hidden risks of carbon destabilization triggered by extreme hydrological pulses. Recently, studies have achieved critical transitions from macroscopic factors to microscopic mechanisms, and from single interfaces to multi-interface coupling. We examined the cascading feedback theories of seasonal rainfall variation on soil organic carbon dynamics in estuarine wetlands. The academic community has gradually clarified its core evolutionary pathways. At the carbon input end, drying-rewetting pulses drive the spatiotemporal decoupling of above- and below-ground primary productivity, reshaping the carbon supply pattern. At the carbon emission end, moisture fluctuations prompt shifts in resource acquisition strategies of plants to regulate autotrophic respiration and reorganize microbial niches to trigger pulsed greenhouse gas releases. Regarding fractions and chemical structure, drying-rewetting cycles accelerate the phase transformation between labile and stable carbon and the selective enrichment of micro-molecules. Finally, from the perspective of sources and stability, the deterioration of physicochemical stabilization mechanisms, such as microaggregate breakdown and the failure of iron-carbon complexation, acts as the core pivot triggering fundamental carbon source-sink transitions. Although the theoretical outline has emerged, there are significant deficiencies in quantitative analysis and mechanistic coupling of multi-interface ecological processes. Future research should focus on: identifying the critical drought thresholds that trigger carbon source-sink abrupt shifts, unraveling the response mechanisms of deep carbon dynamics to plant carbon allocation decoupling, exploring the ecological buffering roles of multi-trophic interaction networks in maintaining carbon stability, and developing process-based predictive models coupling fine carbon fractions with multiple stabilization mechanisms.

Coastal wetlands are biogeochemical hotspots for the global sulfur cycle. Driven by the combined influences of hydrodynamic processes, multiple material inputs, and pronounced redox gradients, sulfur undergoes complex valence-state variations and speciation transformations in coastal wetlands. We systematically summarize research advances on sulfur cycling in coastal wetlands in nearly four decades, with a focus on the major forms and classification of inorganic sulfur pools. We discuss the impacts and underlying mechanisms of microorganisms, plants, key physicochemical environmental factors, and anthropogenic disturbances on inorganic sulfur transformation and mobility. We further examine the role of sulfur cycling in mediating carbon source-sink processes in coastal wetlands. Firstly, sulfate reduction regulates sedimentary carbon transformation by mediating anaerobic organic matter mineralization. Secondly, organic matter sulfurization, sulfur oxidation-driven chemolithoautotrophic dark carbon fixation, and iron-sulfur mineral formation and associated mineral protection of organic carbon collectively contri-bute to carbon stabilization and burial. Thirdly, sulfur cycling regulates coastal wetland greenhouse gas budgets by modulating the production and emission processes of CO₂ and CH₄. Future research should focus on the sulfur cycle response mechanisms under global change, multi-element coupling processes, and the construction of sulfur-carbon cycle models. This review would provide a scientific basis for enhancing ecosystem management and blue carbon sequestration in coastal wetlands and supporting the restoration and maintenance of ecosystem services under ongoing global climate change.

Filter-feeding bivalve aquaculture enhances the transport of particulate organic carbon (POC) to the seafloor through filtration, selective feeding, and biodeposition, thereby altering the sources, fluxes, and reactivity of sedimentary organic carbon (SOC). However, aquaculture-induced organic carbon inputs are characterized by “high input, high reactivity, and high uncertainty”. After entering the sediment, a fraction is returned to the water column through microbial remineralization, resuspension, and lateral transport, whereas another fraction is stabilized through rapid burial, mineral protection, and microbial reworking, which is ultimately retained through long-term burial. Whether enhanced POC sedimentation results in a net sedimentary carbon sink therefore depends on the partitioning of newly added organic carbon among remineralization, physical export, transient preservation, and long-term burial. The current debate centers on whether increased POC sedimentation or SOC accumulation in surface sediments can demonstrate a contribution of bivalve aquaculture to sedimentary carbon sequestration. Assessments based solely on sedimentation fluxes or SOC increments may overestimate the carbon-sink effect, whereas an exclusive emphasis on remineralization losses may underestimate the contribution of stabilization and long-term burial. Following the process sequence of “input-remineralization-stabilization and burial”, we synthesized the transformation processes and controlling mechanisms of SOC in filter-feeding bivalve aquaculture systems and clarified the conditions under which aquaculture-induced organic carbon inputs could generate a long-term positive increment in SOC burial and thereby constitute a net sedimentary carbon sink. Future studies should strengthen synchronous observations and coupled accounting of these processes across relevant timescales, quantify SOC burial efficiency and its influencing factors among different bivalve species, farming practices, and environmental conditions, and establish a framework for assessing sedimentary carbon sinks in bivalve aquaculture.

Mangrove wetlands are important ecotone linking terrestrial and marine ecosystems and provide essential habitats that maintain coastal fishery resource diversity and ecosystem functioning. With increasing global concerns on mangrove conservation and restoration, the protection and sustainable utilization of mangrove fishery resources have become important research topics. We summarized the composition and community structures of fishery biological resources in mangrove wetlands, elucidated the ecological roles of mangrove habitats in sustaining fishery resources, reviewed methodological advances ranging from traditional surveys to emerging technologies such as environmental DNA (eDNA) and remote sensing, and explored the effectiveness of sustainable utilization models, including ecological aquaculture and resource management strategies. Research on mangrove fishery resources has undergone several important transitions, including a shift from resource surveys and production assessments to investigations of ecological functions and ecosystem services, from local community studies to analyses of landscape connectivity and regional ecological processes, and from conventional surveys to integrated monitoring frameworks incorporating eDNA, remote sensing, and intelligent technologies. Resource utilization strategies have shifted from simple exploitation toward the coordinated development of ecological conservation, ecosystem restoration, and fishery production. These transitions indicate that mangrove fishery resource research is moving toward a new stage centered on ecosystem integrity and sustainable utilization. However, insufficient long-term monitoring, limited understanding of ecological processes, and inadequate coordination between conservation and resource management remain major challenges. Future efforts should focus on strengthening long-term monitoring and studies associated with ecological processes, to promote the synergistic development of resource conservation, ecological restoration, and coastal livelihoods.

The coastal China seas are a representative marine region jointly influenced by global climate change and human activities. As primary producers in marine ecosystems, phytoplankton community structure strongly influences ecosystem stability. Here, we summarized the spatiotemporal patterns, driving mechanisms, and modeling approaches of phytoplankton community structure shifts in the coastal China seas. Phytoplankton communities exhi-bit pronounced regional heterogeneity. In nearshore waters, eutrophication and changing nutrient regimes have generally weakened diatom dominance and increased the relative abundance of dinoflagellates and small-sized phytoplankton, accompanied by an increased risk of harmful algal blooms. In offshore and shelf-edge regions, ocean warming and enhanced stratification have promoted a trend toward smaller phytoplankton in size classes. Transitional zones are regulated by the combined effects of water-mass exchange, frontal dynamics, estuarine plumes, and upwelling currents, which collectively drive enhanced spatiotemporal variations in phytoplankton community structure. Community succession is driven by interactions among resource supply, environmental filtering, and biological processes, with temperature, light, nutrient stoichiometry, and microzooplankton grazing playing key roles. Current predictions of phytoplankton community dynamics still face challenges, including limited understanding of ecological mechanisms, simplified functional-group representations, difficulties in parameterizing key processes, and inadequate resolution of mesoscale and submesoscale dynamics. Future studies should strengthen long-term observations and mechanism-based experiments, and develop multi-scale coupled models integrating multiple nutrient forms, high-resolution physical processes, and machine-learning approaches to improve predictions of phytoplankton community dynamics and ecosystem responses.

The 7th Mangrove, Macrobenthos and Management Conference (MMM7) was held in Ishigaki City, Okinawa, in July 2026. The conference brought together recent advances in mangrove ecology, blue carbon, biodiversity, conservation and restoration, social-ecological governance, and spatial monitoring. It featured 5 keynote lectures, 81 oral presentations, and 191 poster presentations, of which 96 were also presented as flash talks. This conference provided a major international platform for showcasing the latest developments in mangrove research. Two research teams from China’s mainland presented their mangrove-related research at the conference. We summarized the major advances in global mangrove research by conducting a systematic synthesis of the conference presentations and posters presented at MMM7. Overall, contributions were evenly distributed across four themes: mangrove biodiversity and ecological processes; blue carbon and greenhouse-gas emission; conservation, restoration and social-ecological governance; and spatial monitoring, geomorphology and hydrology. The research focus is shifting from conventional area inventories, carbon-stock estimation, and species records toward multi- greenhouse gases net climate-benefit assessment, functional biodiversity, process-based restoration, remote-sensing- and artificial-intelligence-enabled monitoring, reporting and verification (MRV), and equitable governance. Comparisons among countries and regions indicate that Indonesia, Australia, the United States, China’s mainland, Hong Kong, China, and Taiwan, China have established relatively complete evidence chains linking national or regional MRV systems, long-term process observations, and management applications. Research in South Asia, Africa, and Pacific island countries remains concentrated on closing baseline data gaps, building regional networks, and strengthening community adaptation. The conference revealed an emerging paradigm centered on mangrove systems that are measurable, mechanistically interpretable, restorable, and governable. Future priorities should include long-term integrated observations, multi-pathway greenhouse-gas emission, cross-scale causal inference, functional biodiversity indicators, and community co-designed assessments of multiple benefits.

Vegetation carbon storage is an important indicator for evaluating carbon sequestration benefits of ecologi-cal restoration. While most available studies have relied on static comparisons before and after restoration, direct quantitative evidence remains limited regarding the dynamics in vegetation carbon storage and its temporal response during restoration. Taking the Jiangsu Sihong Hongze Lake National Wetland Nature Reserve as the study area, we constructed a 30 m reflectance series for 2020-2025 using multi-source remote sensing data. Combined with fine classification of eight dominant vegetation communities and type-specific biomass inversion models, we systematically estimated the spatiotemporal patterns of vegetation carbon storage in the reserve and quantitatively evaluated the carbon enhancement effects of 14 ecological restoration projects initiated successively during 2019-2023. The results showed that the mean of the annual baseline vegetation carbon storage during 2020-2025 was 0.18 × 10⁶t C, and the value in 2025 was 0.15 × 10⁶t C. Spatially, vegetation carbon storage exhibited a pattern of being higher in the lakeshore zone and lower in the open lake area. With the project initiation year as the baseline and 2025 as the unified comparison time point, we found that the 14 restoration projects showed a net increase of 1429.25 t C in vegetation carbon storage relative to their baseline years, with an average carbon gain of 2.75 t C·hm^-2. In the key restoration area where the Yangmaozui degraded wetland ecological restoration project Phase II and species conservation-related projects were located, four projects covering approximately 233.33 hm² generated a cumulative carbon increase of 713.12 t C, with an average gain of 3.06 t C·hm^-2. Based on a comparison by restoration duration up to 2025, the projects were divided into four groups of 2, 3, 4, and 5 years. The 2-year group still showed a short-term net decrease in vegetation carbon storage, with a mean change of -1.59 t C·hm^-2. In contrast, the 3-, 4-, and 5-year groups all showed carbon enhancement, with mean net increases of 7.77, 6.01, and 1.79 t C·hm^-2, respectively. These results indicated that the enhancement of vegetation carbon storage by ecological restoration in shallow lake wetlands would not become stable immediately after implementation, but gradually emerge through vegetation recovery and community reconstruction. This study could provide quantitative evidence for understanding the carbon enhancement process of ecological restoration in lake wetlands and offer a scientific basis for restoration performance evaluation and carbon-oriented management.

Soil dissolved organic matter (DOM), as a key component of the active organic carbon pool in wetland, is deeply involved in organic carbon migration, microbial transformation, and carbon release, and plays an important role in wetland carbon turnover and carbon pool stability. However, the spatial distribution and dominant controlling factors of wetland soil DOM at large geographic scales remain poorly understood. In this study, we selected 57 freshwater wetland sites spanning approximately 4000 km from north to south in eastern China. Using spectroscopic methods, we investigated the spatial distribution of soil DOM in eastern freshwater wetlands, and identified the key driving factors by integrating climatic, vegetation, soil physicochemical, and microbial community variables. The results showed that, along the north-to-south climatic gradient, the quantity and humification degree of wetland soil DOM showed a significant decreasing trend, whereas biogenic characteristics increased signifi-cantly. In terms of component composition, humic-like components dominated in colder regions, whereas protein-like components accounted for a higher proportion in warmer regions. Across different wetland types, peatlands had significantly higher soil DOM quantity than other wetland types, with a relatively high proportion of humic-like components. In contrast, lakeshore and riparian wetlands showed stronger biogenic characteristics and higher proportions of protein-like components, indicating that the specific hydrothermal environments and carbon input-output pathways of different wetland types may jointly regulate DOM accumulation and component differentiation. Results of hierarchical partitioning analysis further showed that DOM quantity, characterized by dissolved organic carbon, was mainly influenced by soil resource conditions, such as total carbon, total nitrogen, and available nitrogen, with a relative contribution of 61.5%. DOM humification degree and biogenic characteristics were mainly driven by bacterial communities and climate, with relative contributions of 34.4% and 35.7%, respectively. Spatial variation in DOM component composition was mainly regulated by climate, with a relative contribution of 64.0%. Overall, the spatial pattern of soil DOM in freshwater wetlands of eastern China is jointly shaped by the synergistic effects of climate and wetland type, while DOM accumulation, optical characteristics, and component composition are driven by different environmental constraints and ecological processes. This study could provide basic data and theoretical reference for evaluating the stability of active carbon pools in wetlands under climate change.

Reclamation aquaculture and invasions by alien species have caused severe degradation of coastal tidal-flat wetlands in China. Mangrove restoration is an important pathway for improving the ecological functions of these tidal flats. At the northern limit of mangrove afforestation in China (Wenzhou, Zhejiang Province), the effects of mangrove planting on the sequestration capacity and stability of soil organic carbon (SOC) remain unclear. In this study, we analyzed the contents and fractions of soil organic carbon, as well as their association characteristics with iron and aluminum oxides across these different land-use types in Yueqing Bay of Wenzhou as the research area. There were four representative land-use types, including bare mudflat representing the pristine state, Kandelia obovata forest representing mangrove restoration, Spartina alterniflora marsh representing alien plant invasion, and aquaculture ponds representing intensive anthropogenic use. The results showed that: 1) Soils under K. obovata and S. alterniflora had higher contents of SOC, labile organic carbon, and dissolved organic carbon, which ranged from 10.76 to 10.85, 1.122 to 1.125, and 0.311 to 0.312 g·kg^-1, respectively, being significantly higher than those in the bare mudflat and aquaculture ponds. 2) Compared with the other soil types, the K. obovata soil exhibited the highest contents of free iron oxide and free aluminum oxide, reaching (12.11±0.74) and (3.91±0.31) g·kg^-1, respectively. Moreover, the proportions of Fe(Ⅲ) and Al-O in the K. obovata soil reached 82.3% and 93.6%, respectively, which were significantly higher than those in S. alterniflora (66.7% and 57.3%), bare mudflat (65.7% and 81.2%), and aquaculture ponds (65.7% and 84.8%). 3) The molar ratios of carbon to iron (C/Fe) in all soils were below 1, indicating that organic carbon was predominantly retained through association with iron oxides by adsorption. The content of iron-bound organic carbon (Fe-OC) was highest in the K. obovata soil (5.82±1.03 g·kg^-1), which was significantly higher than that in S. alterniflora (2.98±0.43 g·kg^-1), aquaculture ponds (2.95±0.16 g·kg^-1), and bare mudflat (2.21±0.31 g·kg^-1). Our results demonstrated that K. obovata mangroves could enhance the sequestration and stability of soil organic carbon in coastal wetlands by promoting the transformation of iron and aluminum oxides into crystalline forms and forming high-valence mineral-organic complex interfaces.

Tidal creeks are ubiquitous in coastal wetlands and act as key conduits for the transport of nutrients and carbon, making them potential hotspots for greenhouse gas emissions. Nevertheless, the spatiotemporal patterns and controlling factors of greenhouse gas emissions from tidal creeks remains rather limited. To investigate the spatiotemporal patterns of dissolved CO₂ concentration (C_CO2), water-air CO₂ flux (F_CO2), and the environmental controls, we conducted multi-seasonal observations along spatial transects in the main tidal creek within the Shanyutan Wetland of the Minjiang River Estuary, southeastern China. Results showed that the C_CO2 in the tidal creek water ranged from 13.81 to 82.55 μmol·L^-1, with saturations between 105.6% and 548.8%, indicating a supersaturated state. Over the observation period, the tidal creek served as a net carbon source, with CO₂ fluxes varying from 0.03 to 2.93 mmol·m^-2·h^-1. Spatially, the C_CO2 and F_CO2 reached their maxima at the landward end and exhibited a significant decreasing trend along the land-to-sea direction. Dissolved organic carbon, total dissolved organic nitrogen, salinity, and dissolved oxygen were the dominant factors driving the spatial variations of C_CO2 and F_CO2. Seasonally, both C_CO2 and F_CO2 exhibited a distinct seasonal pattern of autumn>winter>summer>spring, which was co-regulated by the seasonal rhythm of external loading, tidal mixing intensity, and the seasonal succession of phytoplankton photosynthesis. This study highlighted the necessity of including tidal creek water carbon emissions in coastal wetland carbon budgets, and underscored the importance of accounting for their spatiotemporal variability to enhance the accuracy and comprehensiveness of regional carbon sink evaluations.

Coastal aquaculture ponds are sources of methane (CH₄), which are non-negligible in regional greenhouse gas budgets. However, the seasonal dynamics of CH₄ production and emission in these systems, as well as the driving factors, remain poorly understood. We explored the seasonal variations of soil CH₄ production potential, water gas interface CH₄ emission flux, and related environmental factors in the cultivation pond formed by natural salt marsh wetland reclamation at the mouth of the Minjiang River, by combining in situ field observations and laboratory cultivation experiments. Results showed that soil CH₄ production potential ranged from 8.78 to 60.44 ng·g^-1·d^-1 and water-air interface CH₄ emission flux ranged from 18.58 to 194.74 μmol·m^-2·h^-1, with seasonal coefficients of variation of 65.8% and 79.9%, respectively. Both variables exhibited seasonal dynamics, following the consistent order of summer>autumn>spring>winter. Soil CH₄ emission peaked in summer and autumn. Soil CH₄ production potential was positively correlated with water-air interface CH₄ emission flux in the aquaculture ponds, indicating that methanogenesis in soils was a key factor regulating CH₄ emission across the water-air interface. Random forest analysis identified soil temperature, dissolved organic carbon, microbial biomass carbon, SO₄^2-, and the gene abundances of methanogens and methanotrophs as the primary drivers controlling the seasonal dynamics of CH₄ production and emission. Overall, CH₄ emissions from coastal aquaculture ponds were dominated by a warm-season peak, with substrate supply as the primary driver. Given that feed input is the main source of organic matter and nutrients in aquaculture ponds, it directly determines the available supply of substrates. In the future, emission reduction should focus on optimizing feed utilization efficiency in summer and autumn, achieving a synergistic improvement of CH₄ emission reduction and aquaculture benefits.

Land reclamation for aquaculture, a major driver of coastal wetland degradation and loss, may influence nitrogen biogeochemical cycling at the region scale. Using acetylene inhibition method, we measured soil denitrification potential (SDP) of a Cyperus malaccensis marsh and the converted earthen aquaculture pond in the Shanyutan wetland of the Miniiang River Estuary, southeastern China. Results showed that, compared with the C. malaccensis marsh, the conversion of coastal wetlands into aquaculture ponds significantly reduced SDP by 61.2% to 71.7%. Total soil carbon, total nitrogen, NO₃^--N, and NH₄⁺-N in pore water significantly decreased by 5.1%, 62.3%, 64.8%, and 53.6%, respectively. The abundances of four denitrification functional genes (nirK, nirS, nosZⅠ, and nosZⅡ) significantly decreased by 85.7%, 92.6%, 91.6%, and 94.1%, respectively, while soil carbon-to-nitrogen ratio significantly increased by 170.4%. Random forest analysis identified soil total nitrogen, porewater NO₃^--N and NH₄⁺-N, and the abundances of nirK and nirS genes as the main factors driving the variation in SDP following the wetland conversion. Overall, reclamation for aquaculture markedly suppressed soil denitrification potential, which in turn diminished the coastal wetland’s natural nitrogen removal and its buffering capacity against exogenous nitrogen loading, consequently raising the risks of nitrogen retention and eutrophication under equivalent nitrogen inputs.

Dissolved organic matter (DOM) represents the most labile fractions of organic matter and drives ecosystem carbon turnover in northern peatlands. However, the effects of increased nitrogen (N) availability on DOM dynamics and the underlying mechanisms remain unclear in northern peatlands. We conducted a 14-year N addition experiment (0, 3, 6, and 12 g N·m^-2·a^-1) in a peatland located in the continuous permafrost region of the northern Greater Khingan Mountains. We examined the effects of N addition on DOM chemical composition and biodegradability in surface water and soil pore water (0-15 cm) using the excitation-emission matrix (EEM) fluorescence spectroscopy and laboratory incubation experiments. We found that dissolved organic carbon (DOC) and dissolved total nitrogen (DTN) concentrations in surface water initially increased and then decreased with the increases of N addition levels. Compared with control, N addition at 6 g N·m^-2·a^-1 had 44.8% and 944.5% higher DOC and DTN concentrations, respectively. In contrast, dissolved total phosphorus (DTP) decreased from 0.20 mg·L^-1 to 0.08 mg·L^-1 with increasing N addition levels. These changes resulted in the increases in the DOC:DTP and DTN:DTP ratios, showing the shift in the nutrient limitation status from N to P. Fluorescence analysis revealed that N addition increased the proportion of relatively labile components such as protein-like components, but reduced the proportion of humic-like components. A 42-day incubation experiment showed that DOM biodegradability initially increased and then decreased with increasing N addition levels. In surface water, N addition at 6 g N·m^-2·a^-1 had the highest DOM biodegradability (30.3%), while in soil pore water, N addition at 3 g N·m^-2·a^-1 had the greatest DOM biodegradability (28.8%). Furthermore, the stepwise regression analyses showed that DOM biodegradability correlated significantly with C quality and nutrient availability in surface water, but correlated significantly with the intrinsic aromaticity in soil porewater. Overall, our findings demonstrated that N enrichment could alter DOM chemical composition and biodegradability in northern peatlands, with implications for understanding C cycling in these ecosystems.

Enclosed aquaculture of hairy crab is an important intensive aquaculture model in freshwater lakes of China. However, there remains a lack of systematic understanding regarding its impact on greenhouse gas emissions and the underlying mechanisms. In this study, we investigated the effects of enclosed aquaculture on CO₂ and CH₄ concentration, diffusion flux and the key regulatory factors by field observation in Yangcheng Lake during the breeding and non-breeding periods, combined with headspace equilibrium gas chromatography method and molecular biological technique. The results showed that Yangcheng Lake was a net emission source of atmospheric CO₂ and CH₄, with significant spatial and temporal heterogeneity in emission intensity. During the breeding period, the CO₂ emission of breeding area, non-breeding area, and drainage area were 152.16, 81.59, and 71.19 μmol·L^-1, while CO₂ diffusion fluxes were 204.32, 61.59 and 41.72 mg·m^-2·h^-1, respectively. CO₂ diffusion flux in the breeding area was significantly higher than that in the non-breeding and drainage area. During the non-breeding period, CO₂ emission in the three areas were 68.86, 59.51, and 101.92 μmol·L^-1, while CO₂ diffusion fluxes were 7.04, 14.51, and 69.74 mg·m^-2·h^-1, respectively, with the drainage area being the strongest emission source. The CH₄ concentration and diffusion fluxes exhibited a pattern of drainage area > breeding area > non-breeding area in both periods, indicating that the drainage area was the area with the highest CH₄ emission throughout the year. Random forest analysis and structural equation modelling revealed that Chl-a, pH, dissolved organic carbon and total nitrogen were key factors regulating CO₂ emission, while CH₄ emission were primarily driven by dissolved organic carbon, NH₄⁺-N and the mcrA/pmoA ratio. In conclusion, hairy crab breeding activities significantly enhanced greenhouse gas emission, and thus the drainage system should be considered in the greenhouse gas assessment.

To identify the spatial variations in the relationship between dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC) in global lakes and the major driving factors, we constructed a “hydrological-nutrient” scenario analysis framework using methods including MissForest imputation, classification and regression trees (CART), and generalized additive models (GAMs). Based on global lake DOC and DIC observational data, we compared the correlations between DOC and DIC under different environmental conditions and analyzed their driving factors. The results showed that the CART model identified two primary environmental segmentation points: an annual rainfall of approximately 457 mm and a total nitrogen concentration of about 1.7 mg·L^-1. Global lakes could thus be classified into four scenarios: humid-high nitrogen, humid-low nitrogen, arid-high nitrogen, and arid-low nitrogen. There were significant differences in the relationship between DOC and DIC across these scenarios. In humid-high nitrogen lakes, a weak positive correlation was found, with water temperature near 17.5 ℃ exhibiting nonlinear response characteristics, suggesting that biological processes and water stratification may jointly influence the DOC-DIC relationship under eutrophic conditions. In humid-low nitrogen lakes, the correlation was near zero, likely due to DOC being primarily influenced by soil organic carbon inputs from the watershed, while DIC was more controlled by groundwater recharge, rock weathering, and carbonate equilibrium processes. In arid-high nitrogen lakes, DOC and DIC showed a significant positive correlation, indicating that evaporation concentration and synchronous solute accumulation may be key drivers of their co-varying trends. In arid-low nitrogen lakes, the correlation was weaker, with a response peak in chlorophyll-a concentration near 15 μg·L^-1, suggesting that lower primary productivity may exert phased effects on the dissolved carbon component relationship. Overall, conductivity, pH, water temperature, soil organic carbon, rainfall, and chlorophyll-a were key factors driving the DOC-DIC relationship across different scenarios. The DOC-DIC relationship in global lakes did not exhibit stable synchronous changes but instead demonstrated clear scenario dependence and nonlinear characteristics, necessitating comprehensive analysis in conjunction with hydrological conditions, nutrient status, and lake physicochemical environments.

We examined the changes in soil nitrous oxide (N₂O) emission from meadow communities with different grazing exclusion periods (1, 4, and 7 years) in Napahai Wetland. We measured soil N₂O flux using the static chamber-gas chromatography method, combined with the logarithmic response ratio (RR_i) to investigate its spatiotemporal variations, and analyzed the impact mechanisms of grazing exclusion duration on N₂O emissions. The results showed that grazing exclusion significantly promoted soil N₂O emissions. The RR_i value increased with the extension of grazing exclusion duration, with the RR_ivalue under the 7-year grazing exclusion treatment being 5.6 times that of the 1-year treatment. Compared with the free grazing control, grazing exclusion significantly increased aboveground and underground biomass (increased by 42.8%-611.9%), and decreased the Margalef, Shannon, Pielou, and Simpson indices of plants (decreased by 15.1%-59.3%). Grazing exclusion significantly increased soil water content, pH, organic matter, total nitrogen, available nitrogen, and microbial biomass nitrogen, and reduced soil bulk density. Soil temperature and inorganic nitrogen (NH₄⁺-N, NO₃^--N) decreased first and then increased with the increase of grazing exclusion duration. The RR_i values of plant biomass, soil water content, pH, organic matter, total nitrogen, available nitrogen, and microbial biomass nitrogen all peaked under the 7-year grazing exclusion treatment (increased by 7.5%-352.2%). Soil N₂O emission flux was significantly positively correlated with aboveground biomass, soil water content, pH, and carbon and nitrogen content, and significantly negatively correlated with soil bulk density and plant diversity. The structural equation model explained 89.2% of the total variation of N₂O emissions, with aboveground biomass (direct effect: 87.5%), soil water content (76.5%), soil nitrogen content (83.4%), and soil pH (48.7%) as the main positive driving factors. In summary, short-term gra-zing exclusion significantly increased soil N₂O emissions in the Napahai meadow wetland by increasing aboveground biomass, soil moisture, nitrogen availability, and soil aeration (reducing bulk density), and such effect continued to increase with the extension of grazing exclusion duration.

Extreme climate events can substantially affect gross primary productivity (GPP) and the stability of carbon sinks of wetlands by altering hydrothermal conditions. However, national-scale evidence on how wetland GPP responds to different types of extreme climate events across China remains insufficient to support adaptive wetland ecosystem management. We focused on wetland vegetation across China and combined sensitivity analysis with partial correlation analysis to quantify the response intensity and dominant controls of wetland GPP to extreme climate events during 2001-2023, and further analyzed the lag characteristics of wetland GPP responses. The distributed lag models were applied to assess the duration and direction of the persistent effects of extreme climate events. The results showed that extreme climate events showed pronounced regional heterogeneity during 2001-2023, with precipitation-related extreme events exhibiting stronger spatial variability than temperature-related extreme events. Nationally, 98.7% of wetland pixel GPP showed an increasing trend (Sen slope β_sen ranging from 2.112 to 18.640 g C·m^-2·a^-2). Among them, the proportion of pixels with significant GPP changes in the northern temperate wetlands accounted for the highest percentage (48.3%) of the wetland area in this subregion. Warm days (TX90p) and consecutive wet days (CWD) significantly increased the GPP detrended anomaly by 18.26 g C·m^-2·a^-1 and decreased by 34.99 g C·m^-2·a^-1, respectively. TX90p was the dominant climatic control over wetland GPP variability (partial correlation coefficient ranging from 0.180 to 0.488). Wetlands in temperate zones and the plateau region were most sensitive to extreme heat events (with sensitivity coefficient β of TX90p ranging from 0.376 to 0.648). By contrast, subtropical and tropical wetlands were more strongly controlled by precipitation-related extremes (with β of the maximum 5-day precipitation amount ranging from -0.516 to -0.366). Wetland GPP responded synchronously to extreme climate events, exhibiting significant negative memory effects for extreme heat and drought events, a significant positive memory effect for extreme precipitation events, and a negative but non-significant memory effect for extreme cold events. These findings advanced our understanding of wetland GPP responses to extreme climate events and would provide scientific support for assessing wetland carbon-sink stability and informing wetland conservation under increasing climate-extreme risks.

Vegetation restoration following Spartina alterniflora removal is critical for restoring ecological functions in coastal wetlands. However, how different restoration patterns influence soil bacterial communities remains insufficiently understood. In this study, we investigated restored wetlands in the Minjiang River Estuary, southeastern China, including salt marshes dominated by Phragmites australis and Cyperus malaccensis, and Kandelia obovata mangroves of two stand ages (6- and 16-year-old), with unrestored wetlands as control. We used high-throughput sequencing to characterize soil bacterial communities and used structural equation modelling (SEM) to identify the key drivers. The results showed that: 1) Vegetation restoration significantly decreased soil pH relative to the unrestored wetland, but without significant differences among the restoration patterns. Soil total carbon (TC) and total nitrogen (TN) contents showed consistent patterns across restoration patterns, followed the order of C. malaccensis > P. australis ≈ old-growth mangrove (16-year-old) > unrestored wetland > young mangrove (6-year-old). C. malaccensis marsh had significantly higher TC and TN than the unrestored wetland. 2) Vegetation restoration generally enhanced soil bacterial richness and diversity. The number of observed species and the Simpson and Shannon indices differed significantly among treatments, whereas the Chao1 index showed no variation. The observed species richness and Simpson and Shannon indices generally followed the order of old-growth mangrove > young mangrove > P. australis > C. malaccensis > unrestored wetland. Desulfotomaculum was the dominant bacterial genus across all restoration types, with a relative abundance of 9.4%-13.9%. The dominant genera in the unrestored wetland included Alkaliimnicola, Desulfotomaculum, Myxococcus, Pseudomonas, and Thioalkalivibrio, with relative abundances of 7.2%, 9.4%, 5.6%, 7.5%, and 5.3%, respectively. Following vegetation restoration, the relative abundances of Desulfotomaculum and Myxococcus increased by 2.5%-4.5% and 2.6%-4.6%, respectively, whereas those of Alkaliimnicola and Pseudomonas decreased by 3.6%-4.6% and 3.0%-5.5%, respectively. In addition, Thioalkalivibrio and Pseudomonas were more abundant in mangrove wetlands than in salt marshes, whereas Rubrobacter showed the opposite pattern. 3) Soil TC, TN, and microbial biomass nitrogen were key environmental factors associated with the shifts in bacterial community structure. Structural equation modeling further revealed that vegetation restoration primarily altered soil physicochemical conditions, which indirectly drove changes in bacterial diversity and community structure by altering nutrient availability. These findings demonstrated that vegetation restoration after S. alterniflora removal enhances bacterial diversity and restructures bacterial communities by modifying soil environment, with the positive effects of mangrove restoration on bacterial communities becoming more pronounced with increasing restoration age.

Wetland identification is a fundamental prerequisite for the conservation, ecological restoration, and basin-scale water resources management of wetlands. Conventional remote-sensing-based wetland identification mainly relies on surface spectral information and is susceptible to intra-class spectral variability, inter-class spectral similarity, ambiguous boundaries, and seasonal hydrological fluctuations. It is therefore difficult to reveal the hydrological processes underlying wetland formation and maintenance. To address these limitations, we proposed a refined wetland identification method at the hydrological response unit (HRU) scale oriented toward hydrological connectivity. With Weihe River Basin as a validation area, we examined the applicability of this method in a complex watershed environment. Based on daily HRU-scale hydrological fluxes simulated by the SWAT+ model, we calculated nine dimensionless hydrological connectivity indicators and integrates particle swarm optimization with support vector machine to adaptively determine the classification boundaries of river-fed wetlands, temporary water-supply wetlands, groundwater-fed wetlands, and non-wetlands in a high-dimensional feature space. The identification performance was evaluated through reference validation, typical-area image comparison, and hydrological physical consistency assessment. SHAP analysis showed that the evapotranspiration ratio, detention storage coefficient, channel runoff proportion, and lateral flow ratio were the major contributing features in the identification process. Validation results in the Weihe River Basin showed that the receiver operating characteristic curve values (ROC-AUC) of the three wetland types were all close to 0.90, while the overall accuracy of binary wetland-non-wetland identification was approximately 77.8%. Spatially, wetlands were mainly distributed in a network pattern along the main stem of the Weihe River and its tributaries in Shaanxi Province. During 2018-2022, the average wetland proportion was 3.2% in summer and 0.9% in winter. River-fed wetlands accounted for a multi-year average proportion of 72.0%, making them the dominant wetland type, whereas temporary water-supply wetlands and groundwater-fed wetlands occupied relatively smaller proportions. The results indicated that the proposed method could effectively characterize the spatial extent, type structure, and dynamic changes of wetlands at the basin scale. It would facilitate a shift in wetland identification from “spectral discrimination” to “process diagnosis” and thus provide a new technical pathway for the conservation, ecological restoration, and water resources management of wetlands.